Parkinson?s disease (PD) is a neurodegenerative disease with initial motor symptoms attributed to nigrostriatal dopamine depletion, but with increasing awareness of a non-motor symptom complex. Dopamine replacement therapy (DRT) is the most effective treatment for motor features, but fails to effectively treat non- motor features such as cognitive deficits. In fact, DRT can have a deleterious effect on attention and memory, suggesting that these early cognitive deficits may be unreliable markers for future cognitive progression. Mild cognitive impairment (MCI) in PD is common, and can progress at variable rates to dementia, but eventually results in disability and poor quality of life for most patients. Diffuse Lewy body disease is likely the primary pathological substrate for cognitive decline in PD, although reliable biomarkers predicting pathologic burden and risk of conversion to dementia have not been identified. The subgroup of PD patients experiencing visual hallucinations can develop dementia in 2.5 years, with evidence for associated posterior cortical atrophy and hypometabolism. However, since the incidence of visual hallucinations is low, additional sensitive regional markers of visuoperceptual dysfunction may potentially serve as cardinal signs of pathologic density and distribution. The long-term objective of this application is the study of the predictive validity of early cognitive deficits in PD, and identification of functional neuroimaging markers signaling more rapid conversion to dementia. Our hypothesis, based on models of pathologic staging, is that earlier involvement of posterior cortical regions and the dorsal and ventral visual pathways (with or without the presence of visual hallucinations) are reliable signals for cognitive progression. We will pursue the following specific aims: (1) To utilize a prospective longitudinal cohort to evaluate the prognostic value of PD-MCI subtypes in predicting risk for progression to dementia. Serial neuropsychological evaluations will be given across 4 years to mild PD patients to determine if visuoperceptual deficits predict cognitive progression. An exploratory genetic analysis of how SNCA (?-synuclein), MAPT (microtubule associated tau) and APOE (apolipoprotein E) might influence cognitive progression will be conducted. (2) To evaluate the utility of task-activated fMRI as a probe for cognitive progression by investigating altered posterior cortical networks prior to clinical manifestation. An event-related fMRI paradigm of object recognition memory will measure BOLD response in dorsolateral prefrontal, medial temporal and occipito-parieto-temporal regions in PD patients (with and without MCI) at baseline. (3) To determine the anatomical and regional brain activation patterns predictive of cognitive progression. Structural and functional MRI at baseline and annually for 4 years will characterize anatomical and neural network changes predictive of progression. Identification of biomarkers with sensitivity for early prediction and estimation of risk for conversion to dementia will pave the way for effective intervention with neuroprotective therapies during the critical stage when treatment has the greatest impact.